How I Built a Creative Writing Platform With 3 Different Editors

#nextjs #webdev #opensource #writing

I spent the last few months building TaleForge, a creative writing platform that supports three completely different types of storytelling. Here's what I learned.

The Problem

Writers don't just write novels. Some write manga. Some write screenplays. But most writing tools only handle one format — usually just plain text with some formatting.

I wanted to build something that handled all three, with each editor purpose-built for its medium.

The Three Editors

1. Book Editor

A rich text editor built with Tiptap for long-form fiction. Chapter management, word count goals, writing streaks, and exports to EPUB, DOCX, and PDF.

The key insight: writers need distraction-free writing, not more features crammed into the toolbar. So the editor is minimal by default, with advanced features accessible but not in your face.

2. Manga & Webtoon Editor

This was the hardest one. A canvas-based panel editor where you can create pages with different layouts, add speech bubbles, and arrange panels.

For webtoons specifically, I added vertical section editing with drag-to-reorder, because webtoons are consumed vertically (think scrolling on your phone).

The reader component uses IntersectionObserver for lazy loading and saves reading progress per episode.

3. Screenplay & Animation Editor

Proper screenplay formatting (scene headings, action, dialogue, parentheticals) with Fountain-compatible export.

Plus a storyboard view where you can visualize scenes as cards and rearrange them.

Technical Decisions

Stack: Next.js 16, Prisma, PostgreSQL (Railway), deployed on Vercel.

Why Next.js 16: App Router with server components made the data-fetching story much cleaner. Server actions for mutations, RSC for reads.

Why Prisma: The type safety is worth the occasional ORM frustration. Having your database schema generate TypeScript types that flow through your entire app catches bugs before they happen.

Internationalization: 10 languages from day one using next-intl. This was a pain to set up but worth it — about 40% of my traffic comes from non-English speakers.

Offline support: Service worker with a cache-first strategy for the editor. Writers don't always have internet, especially when they're writing at a coffee shop or on a plane.

The Marketplace

Writers want readers. So I built a marketplace where you can publish your work, set pricing (free or paid with Stripe), and build a readership.

The freemium model lets authors offer the first N chapters free and gate the rest behind a purchase. This is how most web novel platforms work, and it aligns incentives — readers sample before buying, authors get rewarded for writing compelling hooks.

What I'd Do Differently

Start with one editor, not three. Building three editors simultaneously was ambitious. I should have launched with just the book editor, validated demand, then expanded.
Invest in SEO earlier. I spent months building features without thinking about how people would find the platform. Content marketing and SEO should start on day one.
Talk to users before building. I built what I thought writers wanted. Turns out, what writers actually want is simpler than you'd expect — reliable auto-save, good export, and no vendor lock-in.